Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
ORIGINAL ARTICLE
Open Access
Labor reallocation and firm growth: benchmarking
transition countries against mature market
economies
Pradeep Mitra1, Alexander Muravyev2,4 and Mark E Schaffer3,4,5*
* Correspondence:
[email protected]
3
School of Management and
Languages, Heriot-Watt University,
Edinburgh EH14 4AS, UK
4
Institute for the Study of Labor
(IZA), Bonn, Germany
Full list of author information is
available at the end of the article
Abstract
This paper uses firm-level survey data to study labor reallocation and firm growth in
the transition countries over 1996–2005, including benchmarking against developed
market economies. The data shows rapid growth of the new private sector and of
the micro- and small-firm sectors, with the size distribution of firms moving towards
the pattern observed in comparable surveys of developed market economies.
Throughout, the regional patterns suggest greater convergence in the transition
countries that joined the European Union in 2004 than in the other, lower-income
transition economies. We also find evidence of Kuznet-Chenery type structural
change across sectors.
JEL codes: L11, O12, P31.
Keywords: Transition; Convergence; Enterprise restructuring; Labor reallocation
1. Introduction
To use the terminology of classical growth theory, the period of socialism put the
transition countries (TEs) off the path of “convergence” with, or “catching up” to, the
mature or developed market economies (DMEs). The gap in per capita output between
centrally planned economies and mature market economies stopped closing in the
1970s and 1980s. The underperformance of the centrally planned economies was a key
factor that triggered the dissatisfaction with central planning and market socialism and
the move from plan to market in Central and Eastern Europe and the former USSR.
The expectation was that the return to the market would put these countries onto a
growth path that would lead eventually to convergence with the mature market
economies operating at the world technological frontier.
Importantly, such catching-up at the aggregate level may be decomposed into
changes (or convergence) at lower levels, in particular, convergence in technology and
productivity and convergence in economic structure. First, the inherent inefficiencies
of central planning implied lags in innovation and diffusion of technology, resulting in
a virtual stagnation in the 1980 and “equilibrium technological gap” (Gomulka 1986).
Abandoning central planning would enable adoption of improved technology and
management practices and enable the resumption of catching-up. Second, at the onset
of the transition, the countries of Eastern Europe and Central Asia looked very
© 2014 Mitra et al.; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribution
License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium,
provided the original work is properly credited.
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
different from the market economies at similar levels of income in terms of endowments
and the structure of economic activity: large but low productivity industrial sectors, small
agricultural sectors that would be more typical of richer, industrialized countries, and
small services sectors. It was expected that the transition process would generate growth
by reallocating factors away from the excessively-large industrial sector and into the
market services that the central planners had repressed (Döhrn and Heilemann 1996;
Raiser et al. 2004).
The process of economic transition, which in the long run manifests itself in
improved macro indicators, should therefore be visible in more immediate indicators at
the micro-level, which allow observing the “mechanics” of the process (Carlin et al.
2013). Indeed, productivity growth and reallocation at the firm level underlie both types
of productivity-driven convergence in transition economies. Inefficient state-owned
industrial firms were expected to downsize; new private firms would spring up, filling
market niches that were neglected by the central planners; firms would adopt proven
Western technology, production methods, and product standards; both new private
firms and privatized state-owned firms would see efficiency improvements driven by
the incentives brought by private ownership.
There is a large literature that deals with various aspects of reallocation at the firm
level in the transition countries. The bulk of this literature focuses on labor reallocation
as the most accurate and reliable indicator of changes at the micro-level. One of the first
such studies is Konings et al. (1994), based on comprehensive data from manufacturing
firms in Poland in the early transition. In that study, the authors find a large drop in net
employment in state-owned firms driven by a jump in the job destruction rate and a
disproportionate job creation in the private sector. In addition, they find that small firms
appear to be considerably more dynamic than large firms.
Similar results indicating that small and new private firms contribute disproportionately
to job creation while state-owned firms are responsible for most of the job destruction are
reported for Bulgaria, Hungary and Romania (Bilsen and Konings 1996), the Czech
Republic (Jurajda and Terrell 2002), and Estonia (Haltiwanger and Vodopivec 2002).
Interestingly, the study by Jurajda and Terrell (2003), which contrasts the experience of
Estonia and the Czech Republic, two countries that had quite dissimilar transition paths
with a rapid destruction of pre-existing firms in the former and a gradual contraction of
the old sector in the latter, finds that the growth of small start-up firms is similar in the
two countries1. The importance of new private business (as compared to the restructuring
of the existing firms) for a successful transition is also emphasized in McMillan and
Woodruff (2002).
Another important conclusion from the literature is that the patterns of employment
growth, job creation and job destruction vary over the transition period: job destruction
dominates job creation in the early transition period, but the magnitude of the two
processes converges at the later stages. In particular, already by 1995 the job reallocation
rates in the CEE countries are similar to those in mature capitalist economies
(about 20 percent) with roughly equal job creation and destruction rates (Davis
and Haltiwanger 1999)2.
More recently, Earle (2012) revisits the magnitude and determinants of the labor
market flows associated with the fall in Romanian industrial employment in the early
transition. On the background of a large drop in aggregate industry employment, with
Page 2 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
declines in each of the disaggregated two-digit sectors, substantial gross flows in both
directions are documented. Workers from the private sector appear to have a greater
probability of exiting their industry, as well as higher probabilities of finding jobs in
services, as compared with workers in state-owned companies.
Jackson and Mach (2009) study reallocation of labor in Poland over the 1988–1998
period. They document job losses in state-owned firms and job creation in new private
firms as dominant trends in employment dynamics, other than exit from the labor
force. A significant share of this involves a spell of unemployment or exit from the
labor force before obtaining a new job in the private sector. Dong and Xu (2009)
examine the patterns and determinants of the labor restructuring process in China
between 1998 and 2002. They document substantial labor retrenchment in the public
sector. In contrast to transition countries of Central and Eastern Europe, China has
experienced a more synchronized pace of job creation and destruction and higher rates
of excess job reallocation, although even there a large chunk of the displaced workers
had extremely long spells of unemployment or permanently left the labor force3.
Despite these numerous contributions, the evidence on labor reallocation and firm
growth – defined as either employment or sales growth – in the transition countries
remains fragmented and incomplete. In particular, there are lacunas in terms of both
country coverage and time coverage, and more importantly in terms of comparing
different transition countries with each other as well as with mature market economies.
While several studies have noted convergence of the transition countries with mature
market economies, e.g. in terms of the size distribution of firms (Jurajda and Terrell
2003) and rates of job destruction and creation (Faggio and Konings 2003), the
evidence of such convergence remains scarce.
In this paper we use cross-country firm-level data to analyze the convergence process
in the transition economies. We ask several research questions. What are the patterns
of labor reallocation and firm growth in the transition countries? To what extent are
the differences in the employment growth rates observed across regions due to the
differences in the endowments of the respective economies –specifically, ownership,
sectoral distribution, and size of firms? Or do they stem from different relationships
between these characteristics and firm growth across the regions? For example, we
expect that “traditional” (state-owned and privatized) firms will contribute less to
employment growth than the new private sector. The sectoral distribution of employment
across traditional and new private firms is different in the EU8 countries4 and in the CIS
countries. How much of the difference in the employment growth rates between the two
groups of economies should we attribute to having different employment shares of
traditional firms? And how much should we attribute to the fact that the employment
growth rates of traditional firms are different in these two groups of economies, reflecting
different progress in enterprise restructuring?
To address these issues, we take advantage of the data from the Business Environment
and Enterprise Performance Surveys (BEEPS) which are large-scale surveys of firms that
have been implemented in the transition countries since the late 1990s. In particular, we
use a series of three snapshots of virtually all transition economies in 1999, 2002 and
2005. The first year of the BEEPS surveys is 1999, and it happens also to be the first postfinancial crisis year, when the transformational recession is more or less over, and growth
starts across the region. We are able to analyze six years of change, convergence and
Page 3 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
growth. In addition, we have BEEPS data from developed and cohesion Europe (Germany,
Spain, Greece, Ireland, Portugal) collected in 2004/05. This allows us to benchmark the
TEs against established market economies in 2004/05, before the disruption of the world
financial crisis that started in 2008.
The major strength of our study is a comprehensive analysis of firm restructuring in
virtually all transition countries and over a long period of time is. In addition, we are
the first, to our best knowledge, to benchmark the patterns of labor reallocation and
firm growth in transition countries against developed economies using compatible data.
The use of decomposition techniques in the analysis of labor reallocation and firm
growth is another distinct feature of our analysis.
The paper is organized as follows. Section 2 describes the survey and data as well as
introduces the country classification we use. Section 3 describes the methodological
approach chosen. Section 4 presents evidence on labor reallocation and firm growth
from the BEEPS surveys. Section 5 concludes.
2. The survey and data
2.1. The survey
Our empirical analysis is based on the data from the Business Environment and
Enterprise Performance Surveys (BEEPS) that have been conducted by the World Bank
and EBRD every three years since 19995. The main purpose of the surveys was to
assess the environment for private enterprise and business development as well as
to study enterprise adjustment and restructuring. The 1999 survey was administered to
approximately 4,100 firms in virtually all transition countries of Eastern Europe and
Central Asia. The sample size was later extended to over 6,500 firms in 2003 and nearly
11,000 firms in 2005 (see Table 1). In addition, the BEEPS were extended in 2004/05 to
include a range of comparator countries from Western Europe and East Asia
(Germany, Greece, Portugal, South Korea and Vietnam were covered in 2004 and
Ireland and Spain in 2005).
The BEEPS were originally planned as a repeated cross-sectional survey. However,
starting from the third wave (2005), some of the previously sampled firms were
re-interviewed, which results in a (restricted) panel sub-set of the data6. Although the
BEEPS survey instrument has been modified each time it was implemented, the range
of questions that remained consistent across surveys is substantial, especially in 1999,
2002, and 2005. In 2008, there was a significant change in the survey instrument which
makes more recent data not fully comparable with the data from the earlier waves.
Survey samples were constructed by random sampling from a national registry of
firms or equivalent with oversampling of some additional categories of firms to ensure
reasonable subsample sizes. The firms covered are drawn from industry and services7,
and, like the population of firms in countries around the world, are mostly SMEs (see
Table 2). A majority of the firms surveyed (60% in 1999, rising to 75% in 2005) are new
private sector firms, i.e., they were private from the point of startup. Privatized firms
make up about 15-25% of the sample, and the remaining 10-15% were state-owned
enterprises (SOEs). The shares of both privatized and SOEs in the BEEPS transition
samples have been falling over time. The samples from the non-transition market
economy comparator countries include very few SOEs and privatized firms, which
means that our benchmark is the market economy private sector.
Page 4 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Page 5 of 22
Table 1 BEEPS composition by country (number of firms sampled)
Country
1999
2002
2005
Total
Albania
163
170
2004
204
537
Armenia
125
171
351
647
Azerbaijan
137
170
350
657
Belarus
132
250
325
707
Bosnia & Herzegovina
192
182
200
574
Bulgaria
130
250
300
680
Croatia
127
187
236
550
Czech Republic
149
268
343
760
Estonia
132
170
219
521
Georgia
129
174
200
Germany
1197
West
811
East
386
Greece
Hungary
503
1197
546
546
147
250
147
250
Kyrgyzstan
132
173
202
507
Latvia
166
176
205
547
Lithuania
112
200
205
517
Macedonia
136
170
200
506
Moldova
139
174
350
663
Poland
246
500
975
1721
Ireland
Kazakhstan
Korea
610
1007
501
501
585
982
598
Portugal
598
505
505
Romania
125
255
600
980
Russia
552
506
601
1659
250
300
550
Slovakia
138
170
220
528
Slovenia
125
188
Serbia & Montenegro
Spain
Tajikistan
223
536
606
606
176
200
376
Turkey
150
514
557
1221
Ukraine
247
463
594
1304
Uzbekistan
126
260
300
686
4104
6667
10762
24879
Vietnam
Total
500
3346
500
An important sampling restriction in the BEEPS is that only firms that have been in
business for at least three full years are considered. For example, the 2005 wave
excludes firms that were established in 2002 or later. This restriction allows to include
in the questionnaires a set of 3-year retrospective questions that help in tracing changes
in firm dynamics8. Its main disadvantage is that the survey may sample “better” firms,
those able to survive for at least three years. Such a survivor bias may be particularly
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Page 6 of 22
Table 2 BEEPS composition by firm size
Firm size
1999
2002
2004
2005
Total
Micro
1093
2241
1569
4145
9048
%
26.68
33.77
46.89
38.53
36.43
Small
905
2242
1038
3499
7684
%
22.09
33.79
31.02
32.52
30.94
Medium
1171
1126
362
1873
4532
%
28.58
16.97
10.82
17.41
18.25
Large
928
1027
377
1242
3574
%
22.65
15.48
11.27
11.54
14.39
Total
4097
6636
3346
10759
24838
%
100.00
100.00
100.00
100.00
100.00
Note: Firm size categories are defined on the basis of employment, micro (1–9 employees), small (10–49), medium (50–199)
and large (200+ employees). The table shows the number of firms in each size category as well as the percentage of firms in
the respective category in each wave of the survey.
important for small startups in the private sector, among which the exit rate is
usually high.
In each firm, face-to-face interviews were conducted with the “person who normally
represents the company for official purposes, that is who normally deals with banks or
government agencies/institutions”. In small firms, interviews usually involved one
respondent. In larger enterprises, the principle respondent often had to consult with
accountants and personnel managers, especially in the case when detailed data about
the firm (such as sales and employment) were requested9.
The main strength of the survey, from the point of view of this paper, are the use of a
consistent survey instrument across virtually all transition countries and range of
market economy comparators, and, for the transition countries, over a substantial
period of time. The main weakness of the BEEPS is the consequence of the wide
coverage and finite budgets: the sample sizes for individual countries are relatively small.
Even in the biggest BEEPS round in 2005, most country samples have fewer than 400
firms. In the first BEEPS surveys in 1999, a typical country sample had about 150 firms.
2.2. The data
Due to the change in the survey instrument in 2008, we have decided to use the first
three waves of the survey – 1999, 2002 and 2005. The retrospective questions allow us
to cover almost a decade of the transition, from 1996 to 2005, in almost all the transition
countries of Central and Eastern Europe and the former Soviet Union. In addition, we use
data from developed market economies of Western Europe (Germany, Greece, Ireland,
Portugal, and Spain) collected in a special wave of the survey in 2004/200510. Table 1
shows the composition of the BEEPS by country and year of implementation.
Small sample sizes for individual countries makes analysis at the country level a
problematic endeavor. Indeed, too great a degree of disaggregation in the analysis
would results in systematic differences across countries and over time being swamped
by noise in the data. We therefore aggregate across countries in much of our analysis.
Our aggregation scheme separates countries according to position in Europe at the
time of the most recent BEEPS surveys in 2005, and according to income as of 1999.
We consider the following countries and groups of countries:
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
I. West Germany
II. Cohesion countries (Greece, Ireland, Portugal, and Spain)
III. EU8 (new members as of May 2004) (the Czech Republic, Estonia, Hungary,
Latvia, Lithuania, Poland, the Slovak Republic, and Slovenia)
IV. Lower middle income transition countries (Albania, Bosnia and Herzegovina,
Bulgaria, Macedonia, Romania, Serbia and Montenegro, Belarus, Kazakhstan, the
Russian Federation, and Ukraine)
a. SEE (Albania, Bosnia and Herzegovina, Bulgaria, Macedonia, Romania, and
Serbia and Montenegro)
b. Middle income CIS (Belarus, Kazakhstan, the Russian Federation, and Ukraine)
V. Low income CIS (Armenia, Azerbaijan, Georgia, Kyrgyzstan, Moldova, Tajikistan,
Turkmenistan, and Uzbekistan)
We will sometimes refer to Groups I and II as the “pre-2001 EU” or “developed
market economies” (DMEs), and groups III-V as “transition economies” (TEs).
The income classification uses the standard World Bank scheme based on GNI per
capita in 1999 (see Table 3). The division between I, II, III and IV is based on EU status
as of 2005, but also matches GNI per capita in 1999 almost exactly. The division
between IV and V is from the WB income classification. Croatia is an outlier – located
in South-Eastern Europe, this country has GNI per capita that is substantially higher
than that of the other countries of the region – and is therefore omitted from the
analysis. East Germany as well as Turkey, Korea and Vietnam were also included in the
BEEPS 2004–05 surveys, but these countries are too heterogeneous for aggregation; we
omit them from the aggregation scheme and analysis11.
3. Methodology
In this study, we rely on two principal methods of analysis. First, we use descriptive
analysis of the data to produce a general picture of labor reallocation and firm growth
in the transition and comparator countries. The descriptive analysis is instrumental not
only for highlighting the differences between countries and groups of countries, but
also for identifying the time trends in the key variables of interest.
We then use decomposition analysis of employment growth to produce a more
detailed picture of regional patterns and convergence in firm growth. The decomposition
method, commonly used in labor economics, attempts to attribute the difference in the
dependent variable across two groups of observations into the difference in the explanatory variables (endowments), the difference in the relationship between the endowments
and the dependent variable (coefficients), and a remaining factor which is the interaction
between endowments and coefficients. In particular, a differential can be decomposed
as follows:
R ¼ y1 −y 2 ¼ ðx1 −x2 Þ β2 þ x2 β1 −β2 þ ðx1 –x2 Þ β1 −β2 ¼ E þ C þ CE
where R denotes the raw differential between the means of the dependent variable y
measured for two groups of observations, x is the row vector of the means of the
explanatory variables x1,…,xk, and β1 and β2 the column vectors of the coefficients for
the two groups obtained in the regression analysis12. In the final part of the expression,
E = Endowments, C = Coefficient, and CE = Interaction of C & E. The question that
Page 7 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Page 8 of 22
Table 3 GNI per capita in the BEEPS countries
Country
GNI per capita
1999, Atlas method,
current USD
GNI per capita
2004, Atlas method,
current USD
Rank
in 1999
Rank
in 2004
Change,
1999-2004
Tajikistan
280
280
LI
LI
Kyrgyz Republic
300
400
LI
LI
Vietnam
370
550
LI
LI
Moldova
410
710
LI
LI
Azerbaijan
460
950
LI
LMI
↑
Armenia
490
1120
LI
LMI
↑
Georgia
620
1040
LI
LMI
↑
Turkmenistan
670
1340
LI
LMI
↑
Uzbekistan
720
460
LI
LI
Ukraine
840
1260
LMI
LMI
Albania
930
2080
LMI
LMI
Bosnia & Herzegovina
1210
2040
LMI
LMI
Kazakhstan
1250
2260
LMI
LMI
Bulgaria
1410
2740
LMI
LMI
Romania
1470
2920
LMI
LMI
Macedonia, FYR
1660
2350
LMI
LMI
Russian Federation
2250
3410
LMI
UMI
↑
Latvia
2430
5460
LMI
UMI
↑
Belarus
2620
2120
LMI
LMI
Lithuania
2640
5740
LMI
UMI
↑
Turkey
2900
3750
LMI
UMI
↑
Estonia
3400
7010
UMI
UMI
Slovak Republic
3770
6480
UMI
UMI
Poland
4070
6090
UMI
UMI
Croatia
4530
6590
UMI
UMI
Hungary
4640
8270
UMI
UMI
Czech Republic
5020
9150
UMI
UMI
Korea, Rep.
8490
13980
UMI
HI
Slovenia
10000
14810
HI
HI
Portugal
11030
14350
HI
HI
Greece
12110
16610
HI
HI
Spain
14800
21210
HI
HI
Ireland
21470
34280
HI
HI
Germany
25620
30120
HI
HI
n/a
2620
LMI
LMI
Serbia & Montenegro
↑
Note: data from the World Bank. LI stands for low-income countries with GNI per capita less than 826 (755 in 1999), LMI
stands for lower middle income countries with GNI per capita more than 826 (755 in 1999) but less than 3,256 (2,995 in
1999), UMI stands for upper middle-income economies with GNI per capita more than 3,256 (2,995 in 1999), but less than
10066 (9266 in 1999), and HI denote high income economies with GNI per capita more than 10,066 (9,266 in 1999).
usually comes up is how to allocate CE. In the Oaxaca-Blinder decomposition, it is
allocated along with coefficients, so that Explained = Endowments and Unexplained =
Coefficients + Interaction. However, CE can also be allocated to E, or even divided
between E and C. In what follows we allocate the interaction effect along with the
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Page 9 of 22
coefficient effect. Our decomposition analysis is based on Ben Jann’s (2005) “decompose”
addin for Stata.
We consider 3 sources of differences in growth: ownership, sectors, and size. The
decompositions are performed for the following groups of countries:
Cohesion group versus EU8 member states
EU8 member states versus SEE group
EU8 member states versus CIS group
SEE countries versus CIS economies.
In these comparisons, the first group plays the role of a benchmark (leaders) while
the second group embraces the countries that, according to the transition literature,
can be regarded as convergers or followers.
There are almost no privatized and state firms in our sample of cohesion countries,
and we therefore drop any remaining privatized and state firms from the Cohesion
group. The TE groups retain these. The benchmark category (excluded from the
decomposition regressions) is new private firms. The regressions contain 2 dummy
variables for the remaining ownership categories, privatized and state-owned firms.
They also include 6 dummy variables for sectors. For simplicity of estimation and
interpretation, we do not interact sector and ownership dummies thus assuming that
the sector growth patterns do not vary by ownership. In contrast, size effects in our
specifications can vary by ownership, since we want to separate size effects from ownership effects (for example, new private firms can grow fast because they are small and/
or because they are entrepreneurial). Therefore, we interact size (average employment
over 2002–05 measured in thousands) and ownership to get size-ownership effects for
the TEs.
4. Empirical analysis
4.1. Descriptive analysis
The general patterns evident in the BEEPS data are broadly in accord with the earlier
findings on firm growth. Table 4 shows the balance between growing and shrinking
firms, i.e., the proportion of the sample of firms that is growing less the proportion that
is shrinking in terms of sales (no growth firms are ignored). Here and below, growth is
in real terms and covers the 3 years preceding the survey. Germany was in a period of
macroeconomic stagnation during the period 2001–04, and this is apparent in the table,
Table 4 Balance between growing and shrinking firms (share of sample) by country group
Sales
Country group
1999
2002
Employment
2005
1999
2002
2005
I. W. Germany
0.051
−0.130
II. Cohesion
0.297
0.159
III. EU8
0.321
0.282
0.277
0.054
0.059
0.064
IVa. SEE
0.223
0.267
0.246
0.188
0.221
0.153
IVb. Mid inc CIS
0.098
0.461
0.520
−0.009
0.245
0.223
V. Low inc CIS
−0.035
0.292
0.340
−0.137
0.118
0.208
nb: E Germany
0.171
0.018
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Page 10 of 22
with the share of firms with growing sales exceeding the share of shrinking firms by
only 5 percentage points. The cohesion countries, by contrast, were growing rapidly,
with the number of firms with growing sales exceeding the shrinking share by 30
percentage points. The EU8 look very similar, and also show a moderate slowdown over
the full period covered by the BEEPS surveys. An acceleration in growth is very
apparent in the poorest countries/slowest reformers (middle income and low income
CIS), going from near stagnation in 1996–99 to rapid expansion in 2002–05.
The picture in terms of employment is more muted. The number of firms with
expanding employment in the EU8 has barely exceeded the number of downsizing
firms from the very start of the BEEPS surveys. This share is, moreover, low compared
to that in the cohesion countries. Here we see the first indication of a possible failure
of convergence: evidence of possible stagnating job growth in the new EU members.
We will return to this point below. The pattern in the other regions is quite different:
in SEE, firms expanding employment have markedly outnumbered firms shedding labor
since 1996–99; and in the middle income CIS and low income countries, stagnation in
1996–99 is replaced by large-scale expansion in more recent years.
Part of the explanation can be attributed to continued downsizing of employment
and restructuring of state-owned and privatized firms. By 2005, these firms are sharing
in the output expansion in TEs with new private firms, but are still shedding labor.
Privatized and state-owned firms, especially SMEs, are rare in the cohesion countries
and are almost absent from the cohesion country samples. Table 5 shows that in terms
of sales, state-owned and privatized firms in TEs went from being relatively stagnant in
1996–99 to predominantly expanding 2002–05, and indeed the share of firms with
expanding sales in the latter period was very similar for state-owned, privatized and
new private firms. In terms of employment, however, state-owned and privatized firms
had large shares of downsizing firms during the period spanned by the BEEPS, though
declining over time. This is in contrast to the new private firm sector, which was
expanding employment. These patterns of employment growth have been previously
found in earlier studies based on smaller samples covering one or few transition countries, such as Richter and Schaffer (1996); Earle et al. (1996) and Bilsen and Konings
(1996), among others. What is interesting and new in the BEEPS data is that the downsizing of the traditional firms, according to the BEEPS data, continues in all the regions
even a decade after the start of transition.
Table 6 looks at job growth in the more concrete terms of job creation (JC), job
destruction (JD), net job growth (JG) and job reallocation (JR) rates, thus focusing on
aggregate employment growth13. The picture is rather different from what one finds in
the balance table Table 4, because that table simply counts firms, and the smaller firms
are more likely to be expanding but contribute less to the aggregate growth14. The table
Table 5 Balance between growing and shrinking firms (share of sample) by ownership
type, TEs only
Sales
Employment
Ownership type
1999
2002
2005
1999
2002
2005
State
0.101
0.270
0.323
−0.282
−0.164
−0.100
Privatized
0.054
0.330
0.368
−0.322
−0.071
−0.142
New private
0.219
0.342
0.341
0.221
0.272
0.241
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Page 11 of 22
Table 6 Job reallocation rate (JRR), job creation rate (JCR), job destruction rate (JDR) and
job growth rate (JGR), by country group
Country group
JRR
JCR
JDR
JGR
Obs.
Developed market economies (2004/2005):
W Germany
0.138
0.060
0.078
−0.018
810
Cohesion
0.192
0.125
0.066
0.059
2,150
EU8
0.180
0.074
0.105
−0.031
2,946
SEE
0.224
0.088
0.137
−0.049
1,762
Transition economies (2005):
Mid inc CIS
0.191
0.116
0.075
0.041
2,080
Low inc CIS
0.229
0.140
0.088
0.052
1,944
EU8
0.173
0.071
0.102
−0.030
1,885
SEE
0.207
0.092
0.115
−0.023
1,249
Mid inc CIS
0.190
0.137
0.053
0.084
1,458
Low inc CIS
0.203
0.100
0.103
−0.003
1,285
EU8
0.193
0.062
0.131
−0.068
1,124
SEE
0.167
0.053
0.115
−0.062
639
Mid inc CIS
0.216
0.049
0.167
−0.118
1,039
Low inc CIS
0.221
0.033
0.187
−0.154
750
Transition economies (2002):
Transition economies (1999):
Note: Job creation rate (JCR) is defined as the sum of all employment gains in the expanding firms in the economy
divided by total employment, job destruction rate (JDR) is the sum of all employment losses in the contracting firms
divided by total employment, job reallocation rate (JRR) is the sum of the two (JCR+JDR) and job growth rate (JGR) is the
difference between JCR and JDR.
shows that job growth (defined as JCR-JDR) is higher in the richer TEs in the 1996–99
period, but this reverses by 2002–05, when the poorer TEs have faster employment
growth. The reversal is driven by both job creation and job destruction. Job creation
rates in the EU8 are lower than those in the cohesion countries, and the acceleration in
JC takes the poorer TEs ahead of the EU8 by 2002–05. Job destruction rates are persistently higher in the EU8 than in the cohesion countries, and fall markedly in the poorer
TEs so that by 2002–05, job destruction is less common than in the EU8. Job reallocation
(defined as JC+JD) is, however, fairly constant across time and across groups of TEs.
Again, these patterns are broadly consistent with previous literature suggesting that high
rates of job destruction are typical of the earlier stage of the transition process and level
off over time while job creation rates increase (Haltiwanger et al. 2003).
How do job creation, destruction and reallocation compare in transition and developed
market economies? Previous such comparisons have been made by Konings et al. (1994);
Davis and Haltiwanger (1999); Haltiwanger and Vodopivec (2002), and the papers in the
symposium edited by Haltiwanger et al. (2003). These studies typically show that during
the socialist period and in the early years of transition, gross job creation rates in stateowned manufacturing did not change hugely and were similar to those in the OECD,
while job destruction rates in the state-owned sector following the start of transition
increased dramatically and then decline. New private sector firms, by contrast, show high
rates of job creation, job destruction, with the former predominant especially in the early
phase of transition. It should be noted that such comparisons need to be interpreted with
caution as they were hampered by lack of full compatibility of samples; in particular,
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
studies for TEs have typically used firm-level data, whereas studies of JC/JD in Western
economies have used establishment-level data. In this respect, the BEEPS data offer a
better opportunity for such comparison. The data presented in Table 6 show that for the
later transition period, the job reallocation rate is actually no higher in the TEs than in the
cohesion countries – about 20% – and has been very steady in the TEs.
4.2. Decomposition analysis
The decompositions come in two forms: (a) decomposition of aggregate or total
employment growth, which is obtained by using regressions weighted by average
employment; (b) decomposition of average or firm employment growth, which is
obtained by using unweighted regressions. The former shows aggregate employment
effects and is comparable to the analysis of job creation and destruction presented
above (Table 6). The latter is comparable to much of literature on the growth of firms,
and to our analysis above (Tables 4 and 5). The definitions of growth are the same as
those used for the job creation/job destruction growth rate definitions.
In what follows, we discuss the results of our decomposition analysis using graphical
representation. The Additional file 1 contains additional details regarding the decomposition analysis for the pair of cohesion and EU-8 countries. Full results for all groups
of countries are available in Mitra et al. (2008).
The main results of the comparison of the Cohesion and the EU8 countries are
shown in Figure 1. These results refer to aggregate employment growth. The first two
bars in the figure show the net growth rates in the two groups of countries and their
contributing factors. In particular, for aggregate employment the net growth rate in the
Cohesion group is equal to 6.40% and is largely determined by sectoral differences in
growth rates and employment (which cumulatively amount to a growth rate of 6.69%)
with only a small negative contribution of size effects (-0.30%)15. Similarly, the net
growth rate in the EU8 is equal to -1.64% and is composed of a large cumulative
sectoral effect (10.48%), a large negative contribution of state ownership (-7.43%), a
somewhat smaller negative effects of the privatized sector (-2.64%) and the size impact
of privatized firms (-2.16%)16. The size of firms in the private sector and in the state
sector contribute little to the aggregate employment growth. Overall, the high
Figure 1 Decomposition of aggregate employment growth: Cohesion vs New EU.
Page 12 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
contribution of growth among the private firms in the EU8 (which is considerably
faster than in the Cohesion group) is completely wiped out by the negative ownership
and ownership-size effects of the state and privatized sectors of -11.85%.
Bar 3 of Figure 1 shows contributions of these factors to the differences in the
employment growth rates across the two groups of countries. Because sectoral growth
among private firms is faster in the EU8 group (followers/convergers, poorer group)
than in the Cohesion group (leader, benchmark group), the corresponding difference
appears as negative number in bar 3 of the figure (6.69%-10.48%=-3.79%). Conversely,
state and privatized ownership effects appear as large negative numbers in bar 2, being
drags on growth in followers, and thus as positive numbers in bar 3 (slowing down the
catch-up process).
Finally, bar 4 of Figure 1 shows the decomposition results, which further disaggregate
the effects of ownership, sector and size into endowment and coefficient effects, as
discussed in the decomposition framework. It appears that the presence of state and
privatized firms in the EU8 (endowment of these countries with such enterprises) slows
down the employment growth rate and, consequently, the catch-up process in these
economies (the contribution of these endowment effects to the difference in the
observed employment growth rates is 7.04% and 4.8% respectively). In contrast, the
differences in the sector/size distribution of firms (endowment effect) as well as in
the sector/size growth rates (coefficient effects) contribute to higher growth in the
EU8 countries.
Figure 2 shows the decomposition results for average employment growth for the
same groups of countries. The main difference from the aggregate employment results
presented in Figure 1 is that the difference in the private sector growth rates is
essentially nil. The reason is that the EU8 new private firms, which are growing rapidly,
are small compared to the Cohesion new private firms, and therefore do not generate
as much growth in aggregate. The general pattern is the same, however: the growth
gap between the Cohesion and the New EU countries is more than fully explained by
slow growth of privatized and state firms in the latter group of countries17.
Decomposition analysis for the EU8 group and SEE countries is shown in Figures 3
and 4. Similarly to what we have observed in the comparison of the Cohesion countries
Figure 2 Decomposition of average firm growth: Cohesion vs New EU.
Page 13 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Figure 3 Decomposition of aggregate employment growth: New EU vs SEE.
with EU8 countries, the private sector growth rate (sectoral/size effects) is higher in the
follower group (SEE), regardless of whether one looks at the aggregate (16.53% versus
10.20%) or average (15.29% versus 7.74%) employment growth. The detailed decomposition results presented in the Additional file 1 show that the growth differential between
the new private sectors in the two groups of countries is not a size effect, but instead is
driven mostly by the contribution of manufacturing employment growth in the SEE
group. However, faster growth of private firms in these countries is completely offset (in
the case of aggregate employment growth) or substantially mitigated (in the case of
average growth) by downsizing of state-owned and privatized firms: the contribution
of this sector, including size effects, is -11.84% in the EU8 and -24.06% in SEE (see, bar 2
of Figure 3 and bar 2 of Figure 4)18.
The last bar of Figure 3 shows large coefficient effects of ownership, suggesting an
important contribution of faster growth (or slower decline) of state-owned and
privatized firms in the EU8 compared to the SEE group. This can be interpreted as
more advanced adaptation/restructuring of these enterprises in the former group of
countries, the leaders in economic transition. Also interesting is the effect of the
Figure 4 Decomposition of average firm growth: New EU vs SEE.
Page 14 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
difference in endowments in privatized firms, which also contributes to faster employment growth in the EU8. The major negative contribution to faster growth in that region is the coefficient effect of firm size in the private sector, which may indicate
maturation of new private firms or the exhaustion of growth opportunities due to increased competition (see also the discussion in the next section below on competition).
Overall, we observe a kind of catching-up story: the new private sector boom is further
advanced and slowing down in the EU8 countries, but the downsizing of the state
sector is also further advanced and slowing down. The results for average employment
growth (see Figure 4) are similar, except for much less pronounced ownership effects.
A comparison of the EU8 and CIS countries is shown in Figures 5 and 6. Qualitatively, the results are remarkably similar to what we have found in the comparison of
the EU8 versus SEE countries. In particular, the private sector is growing faster in the
CIS than in the EU8 group, which is to a large extent a consequence of a much faster
sectoral growth (in particular, in manufacturing) rather than a size effect.
Given the remarkable similarities in the last two comparisons, it is of interest to
benchmark the SEE group with the CIS countries. The results of this decomposition
are shown in Figures 7 and 8. The weighted/aggregate employment story is that the
SEE and CIS new private sectors are almost identical; the big difference is the much
bigger downsizing of the SOE and privatized sectors in SEE. The unweighted/average
firm story is that SEE and CIS look similar across all sectors. The observed difference
between the weighted and unweighted stories implies that the SOE/privatized downsizing in SEE is more concentrated in the larger firms. The catching up story evidently
does not hold in the CIS, since inasmuch as it is less advanced in the transition compared to the SEE group, the downsizing of the state sector would have been expected
to have been stronger.
The notable differences between aggregate employment growth and mean firm employment growth that we have observed in the above analysis suggest a closer look at
the size distribution of firms is warranted. An indication of convergence in this exercise
would be an increase in employment in small firms relative to employment in large
firms in the transition economies. Indeed, planned economies had very few small firms,
and the small firm sector would be expected to grow rapidly during the transition
Figure 5 Decomposition of aggregate employment growth: New EU vs CIS.
Page 15 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Figure 6 Decomposition of average firm growth: New EU vs CIS.
period to fill this gap (see, e.g., World Bank 2005 for detailed discussion of the Russian
case). We would therefore expect to find that the size distribution is evolving towards
the pattern of the developed market economies, and that the new EU members have
caught up more than the poorer TEs.
This is indeed what we see in the BEEPS data. Figure 9 shows the distribution of firm
size in West Germany and the cohesion countries, the EU8 countries and the other
TEs, by four broad size categories: micro (1–9 employees), small (10–49), medium
(50–199) and large (200+ employees). Small and micro firms are most prevalent in the
developed market economies of the EU, and least common in the non-EU TEs, with
the EU8 members occupying an intermediate position. Figure 10 shows the size distribution of firms in the new EU8 moving steadily towards the developed market economy pattern of large numbers of micro and small firms, and by 2005 the distribution is
close to that observed in West Germany and the cohesion group. Figure 11 shows the
same pattern in the poorer TEs, but these countries start in 1999 from a position of
Figure 7 Decomposition of aggregate employment growth: SEE vs CIS.
Page 16 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Figure 8 Decomposition of average firm growth: SEE vs CIS.
even fewer small firms, and although the small firm sector grows between 1999 and
2005, in 2005 it is still some distance from the market economy benchmark.
As noted already, another test of the convergence hypothesis is to use data on
reallocation across industrial sectors in the course of transition, where we would expect
to see a Kuznets-Chenery-type pattern19. Raiser et al. (2004), in a study of 20-odd
transition countries, divide total employment into broad sectors (agriculture, industry,
markets services and nonmarket services), and show that employment shares during
the transition have generally moved towards benchmarks calculated from a sample of
market economy comparators: in particular, the share of industry has fallen and the
share of market services has risen in all TEs. These patterns are also evidenced in the
relative growth rates of firm employment and jobs in the BEEPS surveys, but with a twist.
The employment growth regressions for 1999, 2002 and 2005 show that employment in
trade and services firms has grown consistently faster than in manufacturing firms in the
EU8 countries20. The twist is that, for the lower-income TE country groups (SEE and
CIS), the differential switches size and manufacturing firms grow as fast as services firms
in 1996–99 and then faster than services firms in 1999–02 and 2002–05. When we look
Figure 9 Size distribution in 2004–05.
Page 17 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Figure 10 Evolution of size distribution in New EU countries.
at aggregate employment growth (i.e., job growth), however, the pattern is different
for TEs as a whole – net job growth is slower in manufacturing throughout the
period – which is consistent with the findings by Raiser et al. (2004)21. In short, we have
evidence at the firm level of two different Kuznets-Chenery-type patterns. In the higherincome TEs, the lower rate of employment growth in manufacturing relative to services
reflects primarily a convergence to market economy benchmarks driven by industrial
sectors that were “too large” at the start of transition, and market services sectors that
were “too small”. In the lower-income TEs, the observed pattern of relatively higher rates
of employment growth in manufacturing relative to services is consistent with a bigger
impact of the standard Kuznets-Chenery-type pattern in which, as a country develops and
productivity grows, employment in manufacturing first increases and then decreases.
In sum, the picture painted by the BEEPS data is broadly consistent with both the
basic macroeconomic trends in the region, and with previous sectoral and firm-level
studies: following the “transformational recession” (Kornai 1994) of the mid-1990s, TEs
have been growing, and at a faster rate than that observed in the developed market
Figure 11 Evolution of size distribution in Non-EU TEs.
Page 18 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
economies – convergence is under way. The pattern of growth at the country, sectoral
and firm level show more rapid growth in the private and especially new private
sectors, movement in the size distribution of firms towards the pattern of large
numbers of small firms as seen in developed market economies, more evidence of
convergence in the new EU members than in the poorer TEs, and evidence as well of
Kuznet-Chenery type structural change across sectors.
5. Conclusion
The move from plan to market in Central and Eastern Europe and the former USSR
was to a large extent driven by the expectation that the return to the market would put
these countries onto a growth path that would lead eventually to convergence with the
developed market economies operating at the world technological frontier. More than
two decades after the start of the transition process, it is the right time to ask if such
convergence has indeed been taking place. Most existing studies of convergence focus
either on the macro aspect (convergence in terms of per capita GDP) or the micro
aspect (convergence in firm productivity). Our study belongs to the second group and
focuses on labor reallocation and firm growth.
We use data from several waves of the BEEPS exercise; due to a number of unique
features, these data are particularly appropriate for studying the process of convergence
in the transition economies. The BEEPS consist of a series of 3 snapshots of virtually
all transition economies in 1999, 2002 and 2005 and covers random and representative
samples from these countries. In addition, the 2005 survey contains firm level data
from a number of developed market economies, which makes it possible to directly
benchmark TEs against these economies.
Our analysis of firm growth, sectoral changes and changes in size distribution of
firms provides a clear picture of the convergence process. Overall, the BEEPS data
show a faster growth of firms in the TEs compared with the developed market
economies. The pattern of growth at the country, sectoral and firm level shows
more rapid growth in the private and especially new private sectors, movement in
the size distribution of firms towards the pattern of large numbers of small firms
as seen in developed market economies, more evidence of convergence in the new
EU members than in the poorer TEs, as well as evidence of Kuznet-Chenery type
structural change across sectors.
Endnotes
1
More on different paths of labor market adjustment in transition countries see Svejnar
(1999) and Boeri and Terrell (2002).
2
A thorough survey of these and other earlier studies analyzing employment growth,
job creation and destruction in the transition economies is provided in Haltiwanger
et al. (2003).
3
We thank the reviewer for pointing this out. See also Giles et al. (2006).
4
The Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, the Slovak Republic,
and Slovenia.
5
Details about the survey are available at http://ebrd-beeps.com/ (as accessed on
April 20, 2014).
Page 19 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
6
For example, the two-wave panel covering the years of 2002 and 2005 contains
about 1,400 observations. Due to the small number of observations in the panel, we
ignore this feature of the BEEPS in the analysis that follows.
7
A small number of firms in agriculture, fishery, forestry as well as in power
generation in the 1999 BEEPS survey are classified in this paper as manufacturing.
8
The respondents are usually asked to give two values, as of the time of the interview
and 36 months ago. This is, for example, the case of employment. In some cases, they
are asked to provide the growth rate (in real terms) over the last 36 months. This is the
case of sales and other financial variables, which, if recorded in local currencies, would
be very hard to deal with in a cross-country study.
9
See, e.g., the 2005 report on survey implementation available online at http://
ebrd-beeps.com/wp-content/uploads/2013/09/beeps2005r1.pdf (as accessed on April
20, 2014).
10
Note that the retrospective questions only allow us to generate three-annual growth
rates of key variables such as sales and employment. This can lead to potential underestimation of the dynamic adjustment in a given country or sector. For example, a
small fall in employment over three years in a particular firm may be caused by a large
fall in year one, a large rise in year two and a relatively small fall in year three. We
believe, however, that such profiles are very unlikely and the bias from a wider than
usual time horizon is not very strong. Importantly, this bias should be in the same
direction for both transition and developed economies and is therefore of lesser
importance in comparative studies such as ours.
11
We admit that the aggregation scheme may mask some important differences
between countries from the same region. Most strikingly is this in the case of SEE,
where we have “basket cases” like Bosnia and Herzegovina (which relied heavily on EU
funds) and Macedonia (which in the 1990s had a completely immobile labor market
with extremely small job creation, but also with tiny flows between labor market
states – see, e.g., Lehmann (2010)) and countries like Romania and Bulgaria. However,
as already noted, the small sample sizes make it virtually impossible to conduct analysis at
a more disaggregate level.
12
In the regression analysis, the dependent variable is the growth rate of employment,
and the regressors are a set of sector dummy variables, a size variable (log employment),
dummy variables for state and privatized firms, and the interaction of the state and
privatized dummies with the size variable.
13
Job creation rate (JCR) is defined as the sum of all employment gains in the
expanding firms in the economy divided by total employment, job destruction rate
(JDR) is the sum of all employment losses in the contracting firms divided by total
employment, job reallocation rate (JRR) is the sum of the two (JCR + JDR) and job
growth rate (JGR) is the difference between JCR and JDR.
14
This is particularly true of SEE countries. Although the number of firms that
expanded employment in this region in 1996–1999 was far larger than the number of
firms that downsized, the JDR was considerably larger than the JCR.
15
These numbers are the predicted values from the regression of employment growth
in the Cohesion countries on sectoral dummies and firm size variable (the ownership
variables are missing because the Cohesion sub-sample only includes new private
firms).
Page 20 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
16
These numbers are the predicted values from the regression of employment growth
in the new UE members on sectoral and ownership dummies, firm size variable, and its
interaction with ownership variables.
17
State and privatized firms are particularly large in the TEs with the result that their
slower growth has a more deleterious impact on aggregate rather than average employment growth.
18
What is driving the smaller negative contributions of the state and privatized
sectors in the decomposition of average as opposed to aggregate growth is that these
sectors consist of firms which are relatively large and make a bigger negative contribution to aggregates than to means. We can see this by comparing the values for “State”
and “Privatized” in the “Mean” columns in the weighted and unweighted results in the
Additional file 1. In the weighted results, these are the values of aggregate employment
in the sample, i.e., in the New EU sample, SOE + Privatized = 0.360 + 0.255 = 61.5% of
employment; in the SEE/CIS sample, SOE + Privatized = 0.318 + 0.366 = 68.4% of employment. In the unweighted results, SOE + Privatized = 0.078 + 0.092 = 17.0% of firms in the
New EU sample, and = 0.086 + 0.129 = 21.5% of firms in the SEE/CIS sample.
19
See Kuznets (1955, 1965) and Chenery and Taylor (1968), Chenery and Syrquin
(1975). As market economies develop, their structure changes in various ways. In
particular, the share of agriculture in GDP and employment falls and the shares of
manufacturing and services increase. The sources of these changes in the size of sectors
have been modeled by Rowthorn and Ramaswamy (1997) amongst others as driven by
(exogenous) differences in productivity growth across sectors. Convergence by the
former socialist economies in this context would generate growth by reallocating
factors away from the excessively-large industrial sector and into the market services that
the central planners had repressed (Döhrn and Heilemann 1996; Raiser et al. 2004).
20
These and other results discussed in this paragraph are not shown in the paper, but
are available on request from the authors.
21
The explanation for this contrast is as follows: Raiser et al. (2004) and other studies
that have looked at structural change in this framework use shares of total employment,
whereas the faster growth of manufacturing firms relative to services firms in the
lower-income TEs that we report is based on firm level data, and as already noted, the
changes in employment in smaller firms play a larger role in the latter because of
the growth of the small firm sector in TEs.
Additional file
Additional file 1: Notes on the decomposition results.
Competing interests
The IZA Journal of Labor & Development is committed to the IZA Guiding Principles of Research Integrity. The authors
declare that they have observed these principles.
Responsible editor: Jackline Wahba
Authors’ information
The first version of this paper was written when Mitra was Chief Economist, Europe and Central Asia Region, World
Bank and appeared as “Convergence in institutions and market outcomes: cross-country and time-series evidence from
the BEEPS surveys in transition economies” (World Bank Policy Research Working Paper No. 4819). Views expressed are
the authors’ and do not necessarily reflect those of the World Bank.
Page 21 of 22
Mitra et al. IZA Journal of Labor & Development 2014, 3:13
http://www.izajold.com/content/3/1/13
Author details
1
World Bank, Washington, DC, USA. 2St. Petersburg State University GSoM, Volkhovsky per. 3, St. Petersburg 199004,
Russia. 3School of Management and Languages, Heriot-Watt University, Edinburgh EH14 4AS, UK. 4Institute for the
Study of Labor (IZA), Bonn, Germany. 5Centre for Economic Policy Research (CEPR), London, UK.
Received: 10 March 2014 Accepted: 27 May 2014
Published: 06 Aug 2014
References
Bilsen V, Konings J (1996) Job creation, job destruction, and growth of newly established, privatized, and state-owned
enterprises in transition economies: survey evidence from Bulgaria, Hungary, and Romania. J Comp Econ
26:429–445
Boeri T, Terrell K (2002) Institutional determinants of labor reallocation in transition. J Econ Perspect 16:51–76
Carlin W, Schaffer ME, Seabright P (2013) Soviet plus electrification: what is the long-run legacy of communism?
Explor Econ Hist 50:116–147
Chenery HB, Syrquin M (1975) Patterns of Development, 1950–1970. Oxford University Press, London
Chenery HB, Taylor L (1968) Development patterns: among countries and over time. Rev Econ Stat 50:391–416
Davis SJ, Haltiwanger JC (1999) Gross job flows. In: Ashenfelter O, Card D (eds) Handbook of labor economics,
vol 3. Elsevier Science, New York and Oxford, pp 2711–2805
Döhrn R, Heilemann U (1996) The Chenery hypothesis and structural change in Eastern Europe. Econ Transition
4:411–423
Dong X-Y, Xu LC (2009) Labor restructuring in China: toward a functioning labor market. J Comp Econ 37:287–305
Earle J (2012) Industrial decline and labor reallocation in a transforming economy: Romania in early transition.
IZA J Labor Dev 1(2)
Earle J, Estrin S, Leshchenko LL (1996) Ownership structures, patterns of control, and enterprise behavior in Russia.
In: Commander S, Fan Q, Schaffer ME (eds) Enterprise restructuring and economic policy in Russia. EDI/World Bank,
Washington, DC
Faggio G, Konings J (2003) Job creation, job destruction and employment growth in transition countries in the 90s.
Econ Syst 27:129–154
Giles J, Park A, Cai F (2006) How has economic restructuring affected China’s urban workers? China Quart 185:61–95
Gomulka S (1986) Soviet growth slowdown: duality, maturity, and innovation. Am Econ Rev 76:170–174
Haltiwanger JC, Vodopivec M (2002) Gross worker and job flows in a transition economy: an analysis of Estonia.
Labour Econ 9:601–630
Haltiwanger JC, Lehmann H, Terrell K (2003) Symposium on job creation and job destruction in transition countries.
Econ Transition 11(2):205–219
Jackson JE, Mach BW (2009) Job creation, job destruction, labour mobility and wages in Poland, 1988–1998. Econ
Transition 17:503–530
Jann B (2005) Stata Module to Compute Decompositions of Wage Differentials., RePEc/IDEAS/Statistical Software
Components, http://ideas.repec.org/c/boc/bocode/s444103.html. Accessed 20 April 2014
Jurajda S, Terrell P (2002) Enterprise restructuring in transition: a quantitative survey. J Econ Lit 40:739–792
Jurajda S, Terrell K (2003) Job growth in early transition: comparing two paths. Econ Transition 11:291–320
Konings J, Lehmann H, Schaffer ME (1994) Employment growth, job creation and job destruction in Polish industry:
1988–91. Labour Econ 3:299–317
Kornai J (1994) Transformational recession: the main causes. J Comp Econ 19:39–63
Kuznets S (1955) Economic growth and income inequality. Am Econ Rev 45:1–28
Kuznets S (1965) Economic Growth and Structure. Norton, New York
Lehmann H (2010) Macedonia’s Accession to the EU and the Labor Market: What can be Learned from the New
Member States? IZA Policy Paper No. 14
McMillan J, Woodruff C (2002) The central role of entrepreneurs in transition economies. J Econ Perspect 16:153–170
Mitra P, Muravyev A, Schaffer ME (2008) Convergence in institutions and market outcomes: Cross-country and
time-series evidence from the BEEPS surveys in transition economies, IZA Discussion Paper No. 3863
Raiser M, Schaffer ME, Schuchhardt J (2004) Benchmarking structural change in transition. Struct Change Econ Dynam
15:47–81
Richter A, Schaffer ME (1996) Growth, investment, and newly-established firms in Russian manufacturing. In:
Commander S, Fan Q, Schaffer ME (eds) Enterprise restructuring and economic policy in Russia. EDI/World Bank,
Washington, DC
Rowthorn RE, Ramaswamy R (1997) Deindustrialization: Causes and implications, IMF Working Paper No. 97/42
Svejnar J (1999) Labor markets in the transitional Central and Eastern European economies. In: Ashenfelter O,
Card D (eds) Handbook of labor economics, vol 3–4. Elsevier Science, New York and Oxford, pp 2809–2857
10.1186/2193-9020-3-13
Cite this article as: Mitra et al.: Labor reallocation and firm growth: benchmarking transition countries against
mature market economies. IZA Journal of Labor & Development 2014, 3:13
Page 22 of 22